🤖 AI Summary
Existing LLM-driven scientific discovery systems struggle to simultaneously ensure creativity and factual accuracy, often requiring substantial human intervention. To address this, we propose a “deliberate decontextualization” framework: scientific literature is decomposed into atomic keyword units; a keyword graph comprising 180,000 biomedical papers is constructed; and cross-concept associations—previously unexplored—are automatically identified via graph analytics, logical inference, credibility validation, and collaborative large-model scoring. Based on this, we develop Nuri, an inspiration engine, and a concretization pipeline enabling end-to-end autonomous scientific discovery for the first time. Nuri achieves an AUROC of 0.737 in identifying high-impact papers; the pipeline successfully reconstructs core ideas from top-tier journals in over 85% of cases, generating concepts that are novel, falsifiable, and consistent with frontier research.
📝 Abstract
Recent advances in LLMs have made automated scientific research the next frontline in the path to artificial superintelligence. However, these systems are bound either to tasks of narrow scope or the limited creative capabilities of LLMs. We propose Spacer, a scientific discovery system that develops creative and factually grounded concepts without external intervention. Spacer attempts to achieve this via 'deliberate decontextualization,' an approach that disassembles information into atomic units - keywords - and draws creativity from unexplored connections between them. Spacer consists of (i) Nuri, an inspiration engine that builds keyword sets, and (ii) the Manifesting Pipeline that refines these sets into elaborate scientific statements. Nuri extracts novel, high-potential keyword sets from a keyword graph built with 180,000 academic publications in biological fields. The Manifesting Pipeline finds links between keywords, analyzes their logical structure, validates their plausibility, and ultimately drafts original scientific concepts. According to our experiments, the evaluation metric of Nuri accurately classifies high-impact publications with an AUROC score of 0.737. Our Manifesting Pipeline also successfully reconstructs core concepts from the latest top-journal articles solely from their keyword sets. An LLM-based scoring system estimates that this reconstruction was sound for over 85% of the cases. Finally, our embedding space analysis shows that outputs from Spacer are significantly more similar to leading publications compared with those from SOTA LLMs.